Comparative effectiveness and safety of edoxaban versus warfarin in patients with atrial fibrillation: A nationwide cohort study

Background and purposeThe effectiveness and safety of edoxaban 60 mg and 30 mg for stroke prevention compared with warfarin in patients with atrial fibrillation have not been well-described in a nationwide cohort of Caucasian patients treated in standard clinical practice.MethodsWe used Danish nationwide registries to identify patients with atrial fibrillation during June 2016 and November 2018 who were treated with edoxaban or warfarin and computed rates per 100 person-years of thromboembolic, all-cause mortality, and bleeding events using an inverse probability of treatment weighting approach to account for baseline confounding. We used weighted pooled logistic regression to compute hazard ratios with 95% confidence intervals comparing events between edoxaban 60 mg and warfarin users; edoxaban 30 mg was not included in formal comparisons.ResultsWe identified 6451 atrial fibrillation patients, mean age was 72 years and 40% were females. A total of 1772 patients were treated with edoxaban 60 mg, 537 with edoxaban 30 mg, and 4142 with warfarin. The median CHA2DS2-VASc score was similar between warfarin and edoxaban 60 mg with a score of 3 (interquartile range (IQR) 2–4). In the inverse probability of treatment-weighted pseudo-population, the thromboembolic event rate for edoxaban 60 mg was 0.95 and 1.0 for warfarin, corresponding weighted hazard ratio of 1.00 (95% confidence intervals (CI) 0.59, 1.71). Edoxaban 60 mg users were associated with lower rates of all-cause mortality (3.93) compared to warfarin (6.04), with a hazard ratio of 0.64 (95% CI 0.47 to 0.88). The event rates for bleeding were 3.36 and 3.14, respectively; hazard ratio 1.09 (95% CI 0.77, 1.57).ConclusionEdoxaban 60 mg is a safe and effective treatment compared with warfarin for stroke prevention in routine clinical care for Danish (mainly Caucasian) patients with AF, with non-significantly different risks for stroke and clinically relevant bleeding, but lower all-cause mortality.